Implant failure is caused by a variety of causes, which is a subject of interest to many clinicians and researchers. In addition to problems with the implant itself, placement, and loading, dropouts are occurring in relation to systemic diseases. Among them, BP, which is used to prevent excessive bone resorption in osteoporosis and cancer metastasis, is known as a drug that induces osteonecrosis, which may cause implant failure after tooth extraction in the jaw [12, 13]. According to the AAOMS, BRONJ is defined as follows: (1) present or a history of treatment with antiresorptive or antiangiogenic agents, (2) exposed bone or bone that can be probed through an intraoral or extraoral fistula in the maxillofacial region that has persisted for more than 8 weeks, and (3) no history of radiation therapy to the jaws or obvious metastatic disease to the jaws [11, 14,15,16].

BPs might be administered either orally or intravenously, and oral BPs, such as alendronate, are most frequently prescribed for osteoporosis and osteopenia [17]. Intravenous (IV) BPs, such as risedronate, pamidronate, and zoledronate, are not only effective for osteoporosis, but also for the treatment of hypercalcemia, multiple myeloma, metastatic cancer, and as an alternative in patients who cannot tolerate the gastrointestinal effects of oral BPs [6]. Four patients in our study, have received IV risedronate due to its frequent prescribed tendencies in Korean internal medicinist compared with those of other countries [18].

BRONJ, peri-implantitis, osteomyelitis, and osteoradionecrosis of the jaw are different entities and the etiology and the pathogenesis are of different origin. The pathogenesis of BRONJ starts with the fact that when the integrity of the oral mucosa of a patient taking BP is broken due to dental treatment, the microbiofilm formed on it penetrates [19]. According to Greg Wanger et al. [20], there is a bacterial nanowire that shows conductivity by special microorganisms, especially metal-producing bacteria, and it plays a more important role in penetration. The hypotheses of BP-induced bone destruction are, first, on the direct role of bone, and second, on the indirect effect on the permeability of the gingival epithelium [21]. Peri-implantitis, associated with severe biological complication, is defined as an inflammatory disease affecting tissues surrounding the implant and resulting in bone loss and eventually implant failure [22]. Osteomyelitis of the jaw may be induced either by hematogenous origin or by dissemination of local infection due to odontogenic infection or trauma [5]. Osteoradionecrosis of the jaw is defined as a complication of radiation that causes a disruption of vascular supply or avascular necrosis with bone exposure in jaw bones that fails to heal over a period of 3–6 months in the absence of local tumor recurrence [23].

One of the main distinct ultrastructural findings of the affected BRONJ specimen is the presence of a high number of microcracks [11]. Almost in all our ultrastructural findings through SEM analysis, a high number of microcracks were found in the attached BRONJ bone to the implant surface. Microcracks are defined as sharp edges larger than canaliculi but smaller in size compared to vascular canals [14]. In an animal study by Kim et al. [24], microcracks were found in SEM analysis of the BRONJ model in rats. The presence of microcracks can be explained by the fact that the jaw bone receives frequent loads with a high degree of mechanical stress by functional forces such as mastication that may lead to cracks. However, in healthy bone, these cracks are continuously repaired by the detection of osteocyte cells that transmit signals for repair [24]. In De Ponte et al.’s SEM study, the healthy bone showed the presence of bone lamellae parallel to each other and partially overlapping like roof tiles, alternating to bone lamellae with the same architecture, but with opposite orientation. Meanwhile, the BP-treated bone biopsy showed visible extensive and frequent areas consisting of a honeycomb structure, or areas with half-cells of different sizes and irregular boundaries, occasionally, partially overlapping each other [21]. Moreover, in Lee’s study, in the SEM observation using decalcified bone microsections, the normal bone showed interdigitating attachment of dendritic bone matrixes which were tightly arranged with each other. The dendritic bone matrixes were sequela of cytoplasmic processes of osteocytes, which contained organic bone matrixes and remained after the demineralization of the bone. The interdigitating dendritic bone matrixes produced many Haversian canaliculi, whereas the BP-involved bone showed granular bone matrixes which were more compact than the normal bone. The Haversian canaliculi formed between the granular bone matrixes were reduced in number and sometimes obliterated abortively [25]. Our main findings suggest that microcracks are one of the most distinct features of necrotic bone tissues found near the surface of the failed implants removed from the jaw of patients with BRONJ. The main ultrastructural findings of peri-implantitis are reported to be osteoclastic resorption lacunae, with altered osteocyte spaces that could be associated to the inflammatory process and the consequences of the increased loads on the remnant bone tissue [26]. For osteomyelitis of the jaw, bacterial biofilms with mixed species and the presence of resorption pits filled with bacterial biofilms are the distinct features found in the ultrastructural findings [27]. In the osteoradionecrosis of the jaw, microorganisms including rods, spirochetes, and cocci, with rods being the predominant cell were the distinct features in the SEM and TEM analysis [28]. Our study supports the previous findings that microcracks could be the first step in the pathogenesis of BRONJ, where a high number of microcracks in the bone samples from BRONJ were detected while samples from osteomyelitis and osteoradionecrosis did not present any microcracks [14]; therefore, the null hypothesis of this study would be rejected.

Our null hypothesis is rooted in the concept that irrespective of the underlying disease process, changes in the implant surface due to host response, biofilm formation, or altered local conditions could potentially exhibit certain similarities in terms of ultrastructural features. While the primary focus of our study is to explore the specific characteristics of implant surfaces in the context of BRONJ, we also recognize the potential significance of cross-comparisons with other conditions. However, we emphasize that our null hypothesis is based on the notion that certain ultrastructural changes might manifest regardless of the specific disease entity. It is important to note that our study aims to contribute insights into the ultrastructural aspects of failed implant surfaces in the context of MRONJ, and we acknowledge the complexity and variations between different disease entities. Future research could certainly explore further comparisons between these conditions to validate or refute our hypothesis.

According to Paulo et al., following tooth extraction in case of the chronic treatment with bisphosphonates, the inflammatory process leads to a decrease in pH, which favors the release of bisphosphonates from the bone reservoir to the surgical wound. This further inhibits the proliferation of fibroblasts, epidermal cells, and endothelial cells resulting in delayed closure of the mucosal barrier and prolonging the deleterious effects of exposure of the underlying bone to microorganisms [19].

Hoefert et al. [14, 29] evaluated the possible role of microcracks in the pathogenesis of BRONJ and discussed its causal model. In his study, SEM analysis found that 54% of BRONJ showed microcracks. In 82% of cases, inflammatory and connective tissue reactions were seen within microcracks. Only 29% of patients taking the medication without symptoms and 17% of osteoporotic patients showed microcracks, but not in osteomyelitis and osteoradionecrosis [14]. The reason microcracks occur in BRONJ is related to a decrease in bone remodeling induced by suppression of osteoclast function due to BP. If bacteria penetrate the generated crack, it becomes symptomatic ONJ [29]. Microcracks can be considered an “important first step” in the pathogenesis of ONJ [14, 29]. Kwon et al. [30] found that BRONJ occurring a short time after dental implant surgery would be regarded as a surgery-related complication. Kim et al. [24] compared the bony reversal lines seen in BRONJ and osteomyelitis. In this study, immature bony matrices outlined by thick reversal lines in BRONJ are evidence of rapid bone destruction osteonecrosis. These unrepaired microcracks were significantly associated with the development of BRONJ [24]. In our study, as in the above findings, microcracks were observed in the necrotic bone of patients taking BP, and bone resorption lacunae were also observed.

SEM and EDS is an effective tool for analyzing BP concentration in the jaw bone and provides important insight into BP pharmacokinetics and BRONJ. With SEM–EDS microanalysis, assessment and quantification of the presence of different bone types based on elemental analysis of Ca, phosphorous (P), and N were carried out. Four representative mineralization areas were found, considering the relative atomic Ca, P from the inorganic bone components, and N content from the organic bone component [12, 13, 31]. Therefore, for the analysis of necrotic bone and failed implants in BRONJ patients, a more effective research method was carried out in this study compared to the previously used method. The ultrastructural findings of BRONJ and implant surfaces were analyzed through SEM–EDS and TEM analysis. BPs, especially N-BP, mainly bind to hydroxyapatite bone minerals at the site of resorption and are captured in the osteoclast during bone destruction [8]. Therefore, N-BPs inhibit the prenylation of small guanosine triphosphate (GTP)-binding proteins in osteoclasts. This series of processes eventually lead to the loss of osteoclast function due to the destruction of the cytoskeleton. The main target cell of N-BPs is bone-resorbing osteoclasts; thus, numerous bone resorption lacunae on the surface of necrotic trabecular bone can be confirmed in our TEM findings, and this indicates that the bone resorption lacunae that occurred while the bone was alive are still present [8, 32].

In Aoki et al.’s study [11], the numbers of resorption lacunae and the length of the erosion on the bone surface of vital bones adjacent to the necrotic bones were increased, and these values in the necrotic bones were correlated with those of the vital bones in BRONJ. According to Kniha et al. [26, 33], the poor state of the osteoclast organelle shown in TEM findings indicates that it is less active or underdeveloped. This is different from the osteocytes seen in the hardened bone or the osteoblasts seen at the edge of the mineralized bone. In Christian Gross’s [34] study, osteoclast inactivation and high cell-to-cell fusion rate were found in the osteoclast profile of MRONJ, and the presence of giant, hypernucleated osteoclasts cannot be attributed to increased dendritic cell-specific transmembrane protein (DC-STAMP) triggered cell-to-cell fusion alone. Our previous study [35] also found that dendritic cells and titanium particles were seen in the necrotic bone removed with peri-implantitis. Peri-implantitis is an inflammatory response, and macrophage-like antigen-presenting cells (APCs) migrate around the inflamed impeller. The dendritic cell (DC) is a member of the APC and is known to initiate and regulate immune response to foreign antigens [35, 36]. However, in the necrotic bone of our study, dendritic cells are not visible.

The absence of dendritic cell in our specimens can be explained by the finding of Elsayed et al. in vitro study [36] stating that BP, especially N = BP inhibit the differentiation and function of dendritic cell rendering the microenvironment more conducive to bacterial colonization and subsequent osteonecrosis. Taking into account, most patients in our study consumed BP for more than 1 year suggesting the high accumulation of BP may have severely suppressed the differentiation of dendritic cells.

Through SEM and EDS analysis, titanium particles were found all over the implant surface in various studies. Shibli’s [37] SEM analysis showed four different degrees of organic residues, appearing mainly as dark stains. The surface topography showed grooves and ridges along the machined surface similar to that of the control group. Overall, foreign elements, such as Ca, O, Na, C, Si, and aluminum (Al), were detected in failed implants. The implants from the control group presented no macroscopic contamination and clear signs of Ti. Nguyen et al. [12] studied the surface of the removed implant which was examined in a patient with maxillary sinusitis caused by various causes. Among them, SEM findings at the apex of the removed implants in BRONJ patients showed no cells or lacuna on the irregular bone surface.

Noumbissi et al. [38] showed that metal ions are released from titanium alloy dental implants due to corrosion. The presence of the long-term corrosion not only leads to the release of ions into the peri-implant tissue but also a disintegration of the implant that contributes to material fatigue and even fracture of the abutments, implant body, or both. From our recent study [15, 16], Ti, C, and O from EDS analysis are not harmful elements due to the chemical composition of the implant. However, inorganic impurities such as Al, zinc (Zn), Si, and magnesium (Mg), with other elements such as nitrogen (N), F, P, Cl, and Na contribute to the corrosion process.

The various metals used in the alloy used in the implant—copper (Cu), Al, silver (Ag), vanadium (V), and manganese (Mn)—are associated with high cytotoxicity and reduced cell viability. According to Park et al., the following elements are in decreasing cytotoxity: Cu > Al > Ag > V > Mn > chromium (Cr) > zirconium (Zr) > niobium (Nb) > molybdenum (Mo) > commercial pure Ti (CP-Ti) [39]. Currently, biomedical Ti is available in four commercially pure grades (ASTM I-IV) and several alloys, including Ti-6Aluminum (Al)-4Vanadium (V) (Ti6Al4V; ASTM Grade V). For the four grades of Coptic, ISO 5832–2 and F67-13 specify alongside Ti, the maximum elemental mass fractions of nitrogen (N) (max.: 0.012–0.05 mass %), carbon (C) (max.: 0.03–0.08 mass %), hydrogen (H) (max.: 0.0125 mass %), oxygen (O) (max.: 0.1–0.4 mass %), and iron (Fe) (max.: 0.1–0.5 mass %) contents. The Fe and O fractions increase from Grade I to Grade IV Ti and correlate with the enhancement of the hardness, yield, and tensile strengths but a decrease in corrosion resistance. The elemental composition of Grade IV Ti, the most common commercially pure Grade of Ti used in dental implants, is standardized as follows: N: max. 0.05 mass%; C: max. 0.08 mass%; H: max. 0.0125 mass%; Fe: max. 0.5 mass%; O: max. 0.4 mass%; and Ti: balance. No other metal element fractions are specified or limited for CpTi in the respective standards [40]. Au element is mainly found in the abutment or prosthesis of the implant. The connection between the implant fixture and abutment may result in the release of metal ions. Ti behaves differently when connected to different materials; it acts as an anode when connected to a noble metal such as Au, whereas it is considered the cathode when connected to a base metal [41]. Therefore, in this study, a high percentage of Au is doubtlessly due to galvanic current activity from gold abutment or corroded gold prosthesis in the mastication during mastication, not from coating material, and can be regarded as contributing factor for periimplantitis, especially in patients with compromised bone tissue such as patients with BRONJ. Al nanoparticles act on the immune system and affect not only immune organs but also immune cells [42, 43].

Dental implant-related systemic toxicity of Al nanoparticles is not known. However, it seems to induce an inflammatory response of the Schneiderian membrane by locally inducing immune cell dysfunction and abnormal immune-related cytokine behavior [44, 45]. EDS analysis in our study revealed that in addition to the main titanium element, gold, carbon, oxygen, calcium, phosphorus, and silicon elements were found. Furthermore, it was also revealed that sulfur was found, which was considered to be one of the complicated causes of implant failure in BRONJ patients. Arteaga et al. [46] tested Ti in an environment similar to diabetes, and there was also an increase in Al. Guler et al. [47] compared the failed implant surfaces and looked at the differences between implant types. In his study, C, N, Ca, P, Cl, S, Na, and Si were also released from a titanium oxide layer on the implant surface. The sulfur (S) component present on the implant surface may be related to the end products of the microorganisms. In our case, S was detected in SEM/EDS analysis; however, it may be that S in our study is not necessarily due to BPs, but the complex microorganisms.